Summary:**AWS Weekly Highlights: Groundbreaking Services Debut in Early July 2026**Amazon Web Services conti
referrerpolicy="no-referrer"
style="max-width:100%;height:auto;display:block;margin:0 auto;">
**AWS Weekly Highlights: Groundbreaking Services Debut in Early July 2026**
Amazon Web Services continued its rapid‑fire innovation cadence in the first week of July, unveiling a slate of updates that span AI infrastructure, database modernization, security visibility, and compliance automation. The announcements signal AWS’s push to deepen enterprise adoption of generative AI workloads while tightening operational governance across hybrid environments.
**Key Developments**
Amazon SageMaker HyperPod now supports Disaggregated Prefill and Decode (DPD), a architecture that separates the memory‑intensive prefill phase from the token‑generation decode stage. By allowing customers to scale each component independently, HyperPod reduces latency for large‑language‑model inference and cuts GPU utilization costs by up to 30 % in benchmark tests.
Amazon RDS for Oracle added native support for Oracle Database 26ai, the latest release that integrates AI‑driven query optimization and automated indexing. Existing Oracle workloads can now migrate to RDS with minimal code changes, gaining access to AI‑enhanced performance tuning without managing the underlying OS.
AWS Security Hub launched Network Scanning, a continuous‑assessment tool that discovers open ports, misconfigured security groups, and vulnerable services across VPCs. Findings are correlated with existing Hub findings, giving security teams a unified view of cloud‑native and network‑level risks.
Amazon Aurora DSQL entered general availability for Change Data Capture (CDC), enabling near‑real‑time streaming of row‑level changes to Kinesis, Kafka, or custom endpoints. The feature simplifies building event‑driven architectures and reduces the need for custom polling scripts.
AWS Config expanded its managed rule library by 191 entries, covering new compliance frameworks such as ISO 42001 (AI management